Long-term calibration models to estimate ozone concentrations with a metal oxide sensor
Long-term calibration models to estimate ozone concentrations with a metal oxide sensor
复制标题
DOI：
10.1016/j.envpol.2020.115363
复制
发表时间：
2020-12-01
影响因子：
8.9
通讯作者：
Kelly, Kerry E.
中科院分区：
文献类型：
--
作者：
Sayahi, Tofigh;Garff, Alicia;Kelly, Kerry E.
臭氧 (O-3) 是一种强氧化剂，会对健康产生不良影响。低成本 O-3 传感器，例如金属氧化物 (MO) 传感器，可以补充监管 O-3 测量并提高测量的时空分辨率。然而，MO 传感器数据的质量仍然是一个挑战。犹他大学拥有一个低成本空气质量传感器网络（称为 AirU），主要测量盐湖城山谷（美国犹他州）周围的 PM2.5 浓度。 AirU 套件还包含一个低成本 MO 传感器（8 美元），用于测量氧化/还原物质。这些 MO 传感器表现出对 O-3 的出色实验室响应，尽管它们表现出一些传感器内的变异性。通过将 8 个 AirU 放置在两个空气质量部 (DAQ) 监测站，使用 O-3 联邦等效方法进行一年的现场性能评估，以开发长期多元线性回归 (MLR) 和人工神经网络 (ANN) 校准模型来预测 O-3 浓度。六个传感器用作训练/测试集。其余两个传感器作为保留集，用于评估新校准模型在预测其他同类型传感器的 O-3 浓度方面的适用性。还通过最小绝对收缩和选择算子（LASSO）、MLR 和 ANN 模型执行严格的变量选择方法。变量选择表明 AirU 的 MO 氧化物质和温度测量以及 DAQ 的太阳辐射测量是最重要的变量。 MLR 校准模型表现出中等的性能（R-2 = 0.491），而 ANN 对于保留集表现出良好的性能（R-2 = 0.767）。我们还评估了 MLR 和 ANN 模型在校准期后五个月内预测 O-3 的性能，结果显示中等相关性（R(2) 分别为 0.427 和 0.567）。这些低成本 MO 传感器与长期 ANN 校准模型相结合，可以补充参考测量，以了解 O-3 水平的地理空间和时间差异。 (c) 2020 Elsevier Ltd. 保留所有权利。
Ozone (O-3) is a potent oxidant associated with adverse health effects. Low-cost O-3 sensors, such as metal oxide (MO) sensors, can complement regulatory O-3 measurements and enhance the spatiotemporal resolution of measurements. However, the quality of MO sensor data remains a challenge. The University of Utah has a network of low-cost air quality sensors (called AirU) that primarily measures PM2.5 concentrations around the Salt Lake City valley (Utah, U.S.). The AirU package also contains a low-cost MO sensor ($8) that measures oxidizing/reducing species. These MO sensors exhibited excellent laboratory response to O-3 although they exhibited some intra-sensor variability. Field performance was evaluated by placing eight AirUs at two Division of Air Quality (DAQ) monitoring stations with O-3 federal equivalence methods for one year to develop long-term multiple linear regression (MLR) and artificial neural network (ANN) calibration models to predict O-3 concentrations. Six sensors served as train/test sets. The remaining two sensors served as a holdout set to evaluate the applicability of the new calibration models in predicting O-3 concentrations for other sensors of the same type. A rigorous variable selection method was also performed by least absolute shrinkage and selection operator (LASSO), MLR and ANN models. The variable selection indicated that the AirU's MO oxidizing species and temperature measurements and DAQ's solar radiation measurements were the most important variables. The MLR calibration model exhibited moderate performance (R-2 = 0.491), and the ANN exhibited good performance (R-2 = 0.767) for the holdout set. We also evaluated the performance of the MLR and ANN models in predicting O-3 for five months after the calibration period and the results showed moderate correlations (R(2)s of 0.427 and 0.567, respectively). These low-cost MO sensors combined with a long-term ANN calibration model can complement reference measurements to understand geospatial and temporal differences in O-3 levels. (c) 2020 Elsevier Ltd. All rights reserved.